Lower CAGR in leveraged meta strategies — investigation
Completed qm-wtfm · 2026-08-17
Do leveraged meta variants show lower CAGR due to math vs selection vs bug?
Reproduced: meta-max-sharpe base 10.64% vs lev2 10.32% on same 1991-385m window (mechanical 2x would be 21.78%). No display/annualization bug (diff 0.0). Order: sleeve levered path returns before aggregation, post-filter renorm. Selection changes ~31% months (38->10 sleeves). Gap spread across many months (top10 =33% of gap). Cause is combination of window truncation, intentional feasibility filter+renorm, vol drag/daily reset path dependence, missing-month skip. No implementation bug; misleading-but-correct display.
Provenance
Equity vs S&P (aligned, rebased to 1.0)
Aligned window, both rebased to 1.0 at start — Aligned window, both rebased to 1.0 at start for fair comparison · S&P dashed, experiment solid. X = Date, Y = Equity (rebased to 1.0).
Signal driver & thresholds (x-aligned)
Signal timeline (sparse triggers)
Triggers when overlay takes over base strategy
Stats
| CAGR | 0.00% | Sharpe | 0.00% |
|---|---|---|---|
| Sortino | 0.00% | maxDD | 0.00% |
Disposition
completed - correct but misleading behavior clarified on Meta page and detailed in Results entry
Follow-on work
- Evaluate financing-cost-adjusted levered cohorts
Errors & data gaps
Lev windows truncated to ETF validity intersection; weighted-meta window 54 months only; per-sleeve F4 estimated
Follow-up
gh issue create --title "Follow-up Lower CAGR in leveraged meta strategies investigation" --body "Follow-up to qm-wtfm: Do leveraged meta variants show lower CAGR due to math vs selection vs bug?" --label "priority:p2" --label "pipeline:research" --label "agent:ready"